Comparing the signatures between worms and cancers

library(deconstructSigs)
library(BSgenome)
library(BSgenome.Celegans.UCSC.ce11)
library(BSgenome.Hsapiens.UCSC.hg19)
human_genome <- 'BSgenome.Hsapiens.UCSC.hg19'
source('~/Mutation accumulation/dnarepairdefiiencies/plotting_functions.R')
source('~/Desktop/Git/phd/useful_functions.R')

These are the latest COSMIC signatures:

sbs_cancer_signatures <- read.csv('~/Downloads/sigProfiler_SBS_signatures_2019_05_22.csv')[,3:67]
dbs_cancer_signatures <- read.csv('~/Downloads/sigProfiler_DBS_signatures.csv', row.names = 1)
id_cancer_signatures <- read.csv('~/Downloads/sigProfiler_ID_signatures.csv', row.names = 1)

Some SBS signatures:

Some DNV signatures:

plot_id_only(dbs_cancer_signatures[,c('DBS1','DBS2')])

Adjusting the worm trinucleotide context to human genome:

worm.trinucl.counts

    AAA     AAC     AAG     AAT     ACA     ACC     ACG     ACT     AGA     AGC     AGG     AGT     ATA     ATC     ATG     ATT     CAA     CAC     CAG 
6285644 1815116 1728559 3741813 1721313  856618  792574 1475960 1800396  967792  833982 1473191 2015069 1531789 1582131 3745527 2414319 1067544 1142800 
    CAT     CCA     CCC     CCG     CCT     CGA     CGC     CGG     CGT     CTA     CTC     CTG     CTT     GAA     GAC     GAG     GAT     GCA     GCC 
1579669 1235912  596780  694118  838826 1093194  555188  689493  793540 1055154 1155528 1142397 1726417 2684274  860331 1151169 1535469 1123127  696495 
    GCG     GCT     GGA     GGC     GGG     GGT     GTA     GTC     GTG     GTT     TAA     TAC     TAG     TAT     TCA     TCC     TCG     TCT     TGA 
 553468  967223 1211898  692334  586758  853675 1101184  860019 1064808 1814847 2186895 1103474 1052833 2017565 2123980 1215740 1091255 1797486 2125754 
    TGC     TGG     TGT     TTA     TTC     TTG     TTT 
1124999 1234435 1720453 2189361 2681125 2416305 6285503 

Upload worm signatures

load('yoda1/beta_G_M_prior_I_greta_270919_5b_trunc.RData')

Plot some

Now adjust the base sub and dinucleotide subs of the most interesting signatures to the human genome.

Now compare them

cbind(colnames(humanized.sbs)[similar[,1]],colnames(sbs_cancer_signatures)[similar[,2]])

      [,1]               [,2]   
 [1,] "Mechlorethamine"  "SBS3" 
 [2,] "Radiation"        "SBS3" 
 [3,] "Mitomycin"        "SBS3" 
 [4,] "Radiation"        "SBS5" 
 [5,] "EMS"              "SBS7b"
 [6,] "polk.1.mms"       "SBS7c"
 [7,] "EMS"              "SBS11"
 [8,] "Cisplatin"        "SBS20"
 [9,] "xpf.1.cis"        "SBS20"
[10,] "xpc.1.cis"        "SBS20"
[11,] "AristolochicAcid" "SBS22"
[12,] "xpc.1.aa"         "SBS22"
[13,] "EMS"              "SBS23"
[14,] "Aflatoxin.B1"     "SBS24"
[15,] "xpf.1.aa"         "SBS27"
[16,] "Aflatoxin.B1"     "SBS29"
[17,] "EMS"              "SBS30"
[18,] "xpc.1"            "SBS40"
[19,] "xpf.1"            "SBS40"
[20,] "Radiation"        "SBS40"
[21,] "Xray"             "SBS40"
[22,] "Mitomycin"        "SBS40"
[23,] "mlh.1"            "SBS44"

Now plot them side by side

Now let’s compare DNV signatures.

names(diinucleotide.freq.factor) <- unique(dnv.types)

Error in names(diinucleotide.freq.factor) <- unique(dnv.types) : 
  object 'diinucleotide.freq.factor' not found

Need to upload the high-res DNV spectra and extract signatures.

dnv.sigs <- nmSolve(D = dnvY, P = X)

4

Now visualize some

Now adjust for human frequencies and compare

Compare

cbind(colnames(humanized.dbs)[similar[,1]],colnames(dbs_cancer_signatures)[similar[,2]])

      [,1]                       [,2]   
 [1,] "xpf.1.UV"                 "DBS1" 
 [2,] "xpa.1.Radiation"          "DBS2" 
 [3,] "xpc.1.Radiation"          "DBS2" 
 [4,] "bub.3..gt2000..Radiation" "DBS2" 
 [5,] "bub.3..ok3437..Radiation" "DBS2" 
 [6,] "rif.1.Radiation"          "DBS2" 
 [7,] "brc.1.MMS"                "DBS7" 
 [8,] "xpc.1.MMS"                "DBS7" 
 [9,] "csb.1.MMS"                "DBS7" 
[10,] "xpa.1.MMS"                "DBS7" 
[11,] "agt.1.MMS"                "DBS7" 
[12,] "polk.1.MMS"               "DBS7" 
[13,] "ced.3.MMS"                "DBS7" 
[14,] "mus.81.MMS"               "DBS7" 
[15,] "xpf.1.UV"                 "DBS11"

DBS11 has no known aetiology, maybe APOBEC mutations. Is not really similar to xpf-1:UV.

Now let’s look at the indel profiles (these are quite tricky to adjust so I’ll compare them on the worm scale).

colnames(Y)[99:112]

 [1] "D.1.rep"      "D.1.nonrep"   "D.2.5.rep"    "D.2.5.nonrep" "D.5.50"       "D.50.400"     "DI.small"     "DI.large"     "I.1.rep"      "I.1.nonrep"  
[11] "I.2.5.rep"    "I.2.5.nonrep" "I.5.50"       "I.50.400"    

regrouped.id.sigs <- rbind(colSums(id_cancer_signatures[c(3:6,9:12),]), colSums(id_cancer_signatures[c(1:2,7:8),]),
                           colSums(id_cancer_signatures[c(27:30,33:36,39:42),]), colSums(id_cancer_signatures[c(25:26,31:32,37:38,73:78),]),
                           colSums(id_cancer_signatures[c(43:48,79:83),]), 0,
                           0,0,
                           colSums(id_cancer_signatures[c(15:18,21:24),]), colSums(id_cancer_signatures[c(13:14,19:20),]),
                           colSums(id_cancer_signatures[c(51:54,57:60,63:66),]), colSums(id_cancer_signatures[c(49:50,55:56,61:62),]),
                           colSums(id_cancer_signatures[c(67:72),]), 0)
rownames(regrouped.id.sigs)  <- colnames(Y)[99:112]